Trial Vaccine Efficacy

Compute attack-risk vaccine efficacy from two trial arms; negative efficacy is retained, not clipped.

Description

Calculate trial-arm attack fractions, their risk ratio, and vaccine efficacy as relative risk reduction from case and participant counts.

Vaccine efficacy compares the risk of a specified outcome in vaccinated and control groups within a clinical trial. The outcome might be infection, symptomatic disease, or another prespecified endpoint. An efficacy percentage describes relative reduction in that outcome under the trial conditions; it does not describe the fraction of vaccinated people who are completely protected. Effectiveness concerns performance in real-world use.^[1]

An attack fraction is the number of participants meeting the case definition divided by the number of participants in that trial arm. The risk ratio divides the vaccinated fraction by the control fraction. Converting one minus this ratio to a percentage gives the attack-risk efficacy estimate. Matching the outcome definition and observation period makes the comparison interpretable.

Trial Vaccine Efficacy performs this arithmetic from four supplied counts. It reports the two attack fractions, risk ratio, and efficacy percent, retaining negative efficacy when the vaccinated arm has the higher observed risk.

When to summarize two trial arms

Use this tool to check an unadjusted point estimate from a two-arm trial table. Choose the participant counts and endpoint consistently with the study analysis; the tool cannot decide which enrolled or evaluable population belongs in each denominator.

Calculate an attack-risk estimate when arm sizes differ.
Check the relationship between arm-specific risks, risk ratio, and efficacy.
Distinguish a relative efficacy percentage from the absolute difference in risks.

How to enter trial counts and calculate efficacy

Enter vaccinated cases and vaccinated participants for the chosen endpoint and observation period.
Enter control cases and control participants using the same definitions.
Run the calculation and read each attack fraction as a fraction; multiply by 100 to express that fraction as a percentage.
Interpret Vaccine efficacy percent alongside the risk ratio, the absolute risks, and the trial design.

Case counts: Nonnegative integers for vaccinated cases and positive integers for control cases. Each case count must be at most its arm participant count.
Participant counts: Positive integers for both arms. Counts are used directly without adjustment for unequal follow-up or attrition.

vaccinatedAttackFraction = vaccinatedCases / vaccinatedParticipants; controlAttackFraction = controlCases / controlParticipants

Each denominator belongs to its own trial arm.

riskRatio = vaccinatedAttackFraction / controlAttackFraction; vaccineEfficacyPercent = 100 × (1 − riskRatio)

A ratio below one gives positive efficacy, one gives zero, and a ratio above one gives negative efficacy.

json

{
  "input": {
    "vaccinatedCases": 10,
    "vaccinatedParticipants": 1000,
    "controlCases": 40,
    "controlParticipants": 1000
  },
  "output": {
    "vaccinatedAttackFraction": 0.01,
    "controlAttackFraction": 0.04,
    "riskRatio": 0.25,
    "vaccineEfficacyPercent": 75
  }
}

Observed comparison	Result
Zero vaccinated cases with positive control cases	Risk ratio 0 and efficacy 100%.
Equal attack fractions	Risk ratio 1 and efficacy 0%.
Vaccinated attack fraction twice the control fraction	Risk ratio 2 and efficacy −100%.
Zero control cases	Input error: the implementation requires a positive control count.

Missing fields, noninteger counts, counts exceeding their arm sizes, and nonfinite values also produce input errors.

Limitations of an unadjusted efficacy point estimate

The tool supplies no confidence interval, significance test, sample-size calculation, or adjustment for randomization strata and covariates.
Participant counts are not person-time. Different follow-up durations require an analysis that accounts for the observation time.
A computed difference does not establish causation without an appropriate study design and comparable ascertainment.
Zero vaccinated cases can produce 100% efficacy in a finite sample. Negative efficacy is retained as an observed comparison; the calculator does not evaluate its statistical uncertainty.

The selected endpoint matters: efficacy against one outcome should not be presented as efficacy against every outcome or as a prediction of protection for an individual.

Alternative or Complementary absolute-risk summaries

Report the arm-specific risks along with relative efficacy. When the vaccinated risk is lower, their absolute difference can feed a number-needed-to-vaccinate calculation for that same endpoint and time horizon. Household attack rates describe another study context and should not be substituted for trial arms without an appropriate design.

References

Vaccine efficacy, effectiveness and protection — World Health Organization

Similar or alternative tools

Number Needed Vaccinate
Calculate absolute risk reduction and its rounded reciprocal from supplied vaccinated and unvaccinated risks; a positive reduction is required.
Household Secondary Attack
Calculate household and pooled attack percentages from susceptible-contact and secondary-case counts.